Informatica: Data Integration in (and for) the cloud
After talking about data integration being the Achilles heel of cloud computing I had a chat with Informatica, who is not only providing a solution for data integration for the cloud, but is leveraging the cloud to do it.
While we at F5 are focused on tearing down the silos that exist in IT to support the delivery and management of applications both internal and external (SaaS, cloud), Informatica is looking to tear down the silos in the cloud that currently exist as Software as a Service (SaaS) offerings.
Integration, always a painful subject, has become even more cloudy as the need to share data between local and remote applications has increased. Loraine Lawson uses a quote gleaned from the trenches in "Integration Is Thorny Issue for SaaS" to illustrate just how painful integration can be for those involved:
Integration problems – isn’t that always the big brick wall IT hits? As someone recently expressed on the Yahoo SOA Group:
“Whenever anyone say ‘we need to integrate our system with this other system’ people shiver and sweat and hope that they’re not part of that project, because down that path lies madness, ad-hoc or not.”
But will integration be more or less of an issue if you move to the cloud?
There’s no doubt that integration has become a major issue for SaaS companies.
Most organizations use some of form of SaaS, whether it's for travel, payroll, or customer relationship management. The problem is, as it usually is when using third-party applications, integration. The data created "out there" needs to be integrated with data "in here" (and vice-versa) and that process is made increasingly difficult by cloud computing.
Informatica aims to change that by reinventing integration, essentially. Rather than deploying a solution in-house and spending time configuring and orchestrating integration flows between disparate applications, Informatica has simplified and moved the process of defining those integration into the flows. Yes, all its configuration is a service, in the cloud.
Once the configuration is complete - Informatica offers turn-key adapters for Salesforce.com and has plans to expand support to other SaaS platforms - a daemon is built. This executable is subsequently downloaded and deployed locally. The daemon handles replication, synchronization, and coordination of data sharing between applications residing in the local data center and those residing elsewhere, in cloud data centers. Once configured and deployed, the local daemon (agent) handles synchronization and replication of data and manages the flows mapped during configuration. That's quite a different process from traditional integration solutions, which require their own servers, licensing, management, and (always difficult for some reason) configuration.
Customizable turn-key integration is about the best way to describe this model, and it's an interesting use of the cloud to make what is, admittedly, usually a difficult and painstaking process a simple and almost pleasant sounding experience.
Informatica's solution essentially makes SaaS offerings appear to internal applications like an RDBMS. That's a familiar model to most IT and even some more savvy line of business folks who live their lives with an Excel spreadsheet always up and pulling data from multiple sources to provide near-time data on which to make current and future business decisions.
Issues with this model? There's a ton of technical and security issues that are fairly obvious from the get-go. The daemon is communicating with a third-party provider and needs access, credentials to internal and external data sources must obviously be supplied to the configuration system, and there needs to be a system on which the daemon can run. The possibility of integration agents running on desktops or unmanaged servers across the enterprise is certainly an unsettling one, and could disrupt network traffic patterns and introduce performance problems across the organization. Troubleshooting connectivity issues inside the organization, too, would certainly be problematic for a remotely configured integration solution.
Still, every new model of computing and integration comes with its problems, and this one is interesting from the perspective of making the actual process and subsequent deployment a simple one. It certainly meets Informatica's goals of ease of use, reducing the costs associated with integration, and decreasing time to market/deployment. It's certainly less time consuming than writing code specifically to perform cross-organizational integration, and more reliable than using native replication capabilities of the RDBMS in the cloud, when they're even available.
It is a solution to the problem of integration between local and remote (cloud) applications - something that is desperately needed as we continue to explore the ways in which SaaS and cloud computing can be used to improve IT and will certainly become increasingly important as the adoption of cloud computing and SaaS continues to increase throughout the coming year.
You can find more information about Informatica's On-Demand Data Integration offering at their site.